What moral decisions should driverless cars make? | Iyad Rahwan
HTML-код
- Опубликовано: 7 сен 2017
- Should a driverless car kill you if it means saving five pedestrians? In this primer on the social dilemmas of driverless cars, Iyad Rahwan explores how the technology will challenge our morality and explains his work collecting data from real people on the ethical trade-offs we're willing (and not willing) to make.
The TED Talks channel features the best talks and performances from the TED Conference, where the world's leading thinkers and doers give the talk of their lives in 18 minutes (or less). Look for talks on Technology, Entertainment and Design -- plus science, business, global issues, the arts and more.
Follow TED on Twitter: / tedtalks
Like TED on Facebook: / ted
Subscribe to our channel: / ted Авто/Мото
I'm a simple man, I'm only here because I have an assignment on it due next week.
What a coincidence, I’m here for the same reason
A coincidence you say? I am as well:)
@@l.k9874 i am literally writing a written exam about this as We speak.
I hear ya bud.
☝🏾me too
I think that a driver-less car should never swerve and only ever brake. If the brakes are broken or there is a problem there should be a loud continuous beep alerting people about it. The car should always stay on the road no matter what and never swerve onto the path even if it will save lives. when a pedestrian is crossing the road they should be aware of the cars, a pedestrian on a side walk should be safe in knowing a car will never hit them and so will not need to be alert. If cars are designed to avoid larger groups then will people walk in groups to increase their survivability? I dont think the cars should make decisions but rather have a predicable manner that people can learn and therefore avoid. such as braking so people can hear the sound and react.
JamieA242 that makes alot of sense. if you make the cars actions more predictable in such event then it puts the decision making back into the hands of humans
I'm not going to wait at a crosswalk for an hour or more on the off-chance another human will cross with me -- in the same direction, as well -- and hold my hand so as to spook the cars by our size as if it were a puma or bear. Sheesh, we have to be afraid of what's in the car AND the car itself? Do you know that from time immemorial, all paths and roads belonged to pedestrians? Feet made the roads, and roads were made for feet. Children could play in the roads, there were roadside stands where people hawked their wares, and people felt safe enough to take a stroll outside with their friends, children or lovers. Then cars came, so loud it terrified all the horses and deafened people, and boy did they make a stink. Worse yet, if wheels weren't flying off, deadly collisions with pedestrians (especially children) with these solid steel monstrosities were also making the automobile very unpopular. About to lose road rights to the carriage cabbies and horseback populace, the automobile industry vilified the victims to the accidents by calling them "jaywalkers". Back then, a "jay" was a very dirty word akin to the "n-" word today, alluding to dirty mountain hicks. NOBODY wanted to be a "jay", and somehow they got away with making it a law, and that helped them win the road war. In short, people by rights own the roads, not cars. Crossing legally at designated crosswalks that are every eighth/quarter mile apart (or whole mile sometimes!) in a city should not be punishable by death if the car's breaks fail.
this is the problem with driverless cars: You can hack them to kill a person. You can drive them, kill someone and call it's the car's fault. The cars do not work in any society.
That's a pretty good answer
Let me ask YOU something. Do ya want more confidence? Better social skills? EASY tips to start improving your quality of life? Well I make videos on all that!! *Just Head to my channel* appreciat ya
Ok Google, put my car in rampage mode.
*Religion of peace mode
It's called GTA mode good gentleman.
ITS CALLED ISLAM MODE
+Double D Not cool, man.
Savage
The real question is this going to be the end of Russian road rage videos?
OH NO! WE NEED TO DITCH THIS EVIL SELF-DRIVING CARS!!!
This talk was much more thoughtful than I expected.
How about creating a transportation system where driverless vehicles don't need to make those moral decisions because the whole system is designed to minimize vehicle-pedestrian interaction (where they have completely different pathways that don't meet because they are both enclosed and at different heights/positions for example) and the vehicles are "fixed" to the road/way (for example cable-railway/maglev/vacuum tubes/etc.)
G_Guy001 Futurama 👍🏻
that's nice point!
that would take to much space to ever be efficient but is a noble idea
a little bit utopic. i see your reasoning though
Your idea probably cost too much both in implementation and maintenance. Driverless cars seem far more realistic and may even save more lives than your idea.
They eliminate human error, and in the case of mechanical failure a driverless car should be able to avoid casualty the vast majority of the time. Other driverless vehicles might even help in the process, especially if the cars are sending wireless signals to each other.
I found their survey a little flawed - it asked loads of questions about different age ranges of pedestrians, and even different types of people (ie fit, athletic person vs criminal, as if the car is going to be looking at pedestrians thinking "now that looks like a criminal, I'll kill him before I kill that honest jogger"). I made the decisions I made in the survey based on a simple pair of rules:
1. Avoid hitting pedestrians if at all possible
2. If it is impossible to avoid pedestrians then take the simplest path (ie don't swerve)
And afterwards it told me that I valued old people more than young etc etc etc.
I think problem is that if you have closed questions you can have a lot of participants, but you have to be careful how you interpret the results. If you make an interview then you know all, but instead of handred thousand you will have thousand participants at best.
Very cool lecture, we need more talks from Iyad Rahwan. Just awesome!
I can already foresee my future where my children will ask me what was it like to drive a car.
General S. Patton
Unless you die before you have children.
I am sure it would still be possible to drive recreationally.
Hello people from the future
Baginka Man Hello back.
hello to both of you
Your message is in the wrong place. This is the present.
What up bro
1) A driverless car should realize when the brake pads are worn to the point they need replacing and take itself out of service before they are completely gone. Even driving itself to the service center. 2) If the brakes start to fade, it should pull itself over safely. 3) If the brakes completely fail, the car should flash it's headlights (left/right like a LEO car), emergency flashers and blow the horn steady to warn pedestrians and human driven cars. At the same time, it should downshift itself harshly to try and slow itself down as much as possible, then shift into reverse and then park to stop itself even if it damages the transmission. *It should also be transmitting a warning other driverless cars of it's failure.* 4) Have the tires scrub the curb or sideswipe others cars to slow itself down before hitting a solid object. - Can you program that?
A self-driving car must always preserve the lives of those onboard, because otherwise no-one in their right mind would voluntarily get on board something that would potentially kill him/her. Yes, this means making the decision of running over 5 kids instead of killing the driver. It is what I would do in a similar situation (brake failure, etc. - running over people instead of driving to a brick wall for instance). Only valid exception in some situations might be if only animals are on board the car.
Self-driving cars aren't fully analogous to say, public transports because the drivers of public transports are human, and you can generally count on them having the same self-preservation instinct as you do. A machine has no such thing, and is perfectly willing to terminate itself so as to not cause harm to other humans.
The real dilemma with self-driving cars is this - unless you have complete access to all of the software and hardware options to the lowest level possible on the car, you cannot be certain as to whether or not the car will choose to kill you in an emergency (think a lower level hardware setting overriding an option of self-preservation selected on the car's computer user interface). But having complete access to the software of the car would also mean that people could (at least definitely way more easily) subvert the programming in nefarious ways, for instance make them drive over people on purpose. Thus no such access will ever be granted to people; ergo, you can never be 100% certain if a self-driving car will potentially choose to kill you, or not.
This is the reason why I will never get on board, own, or "ride" on a self-driving car, ever.
Doakes That is an interesting point. We know the manufacturer is motivated to maximize sales. Sometimes they even take shortcuts to make an extra buck. (Think VW diesels or the Ford that blew up - I think it was Pinto - or the Chevy Vega roll overs.) Given this we can expect them to do as you suggest and program to preserve those in the car. But I don't think we would see consumer access to the code.
Hirvieläin in the future you won't be able to own a Manuel driver car. It would essentially be illegal. However, the car should hold priority to it's passengers. If it can prevent the death of the bystander and the passenger it will do it. It can weigh the percentages in a split second of what decision to make
Most things have the potential to kill us, and yet we choose to still do them; roller coasters, knives, white water rafting, sleeping on a bunk bed. Death is a possibility in each moment, so we need to make the more rational choice and minimize overall suffering. Stepping outside of the debate on how these cars should be programmed, I think this debate is the bike-shed effect all over again (google it); we will all be much safer due to having these cars period, so whatever ethics we decide to program into them will ultimately be trivial by comparison.
Henry Hui That's internet-troll level of stupid
Henry Hui As a purchaser of the self driving car, you should take responsibility for your car and thus, should be sacrificed first in the VERY unlikely case of a "some1 dies" situation. Your life is not more valuable as those of bystanders, so the tiebreaker is that it was your car and you should be sacrificed.
Tbh, people who think like you are the reason why the earth is in the situation it is. No logic, just self-preserving emotion. You are not special, no one is.
I feel that roads should just be considered dangerous areas. The car shouldn't drive into a pavement to avoid damage but should stay on the road. When crossing a road you should be aware of the dangers.
As I frequently encountered articles about self-driving cars, I wondered who should be responsible for hitting pedestrians and causing death. Personally, I think human drivers should be given the choice of what the car should do. Through this, I think drivers can be responsible for the consequences they have committed. Therefore, I think legal regulations on self-driving cars should be strengthened. If people comply with legal regulations, accidents will not occur.
i think i must use this solution for my thesis :(
This reminds me of the trolley prob - op, he just mentioned that. Okay, nevermind.
I dont really think it needs to make any decisions at all, it just needs to get to its destination without any accidents and attempt to avoid the best it can if something occurs, which is not much different than a human driving, a human having to make a choice of who lives or dies in an automobile accident is rare, insanely rare, as it happens so fast there really is no time to contemplate decisions like that.
this to is also maybe a good question to ask, but not as important as it seems, as car accidents are predominantly, of coarse, human error, and that human error is usually caused by things like texting, or lack of sleep, or toxcicity. cars dont do drugs, cars dont get tired, cars dont get drunk, cars dont text or talk on the phone, cars dont look down at the radio, cars dont need glasses or hearing aids, cars dont get old age......this doesnt mean mistakes wont happen, but it already would be FAR FAR less mistakes than humans make, saving probably millions of lives just on its own.
There is no time for a human to make those decisions, but computers can have access to all the aggregated traffic data ever recorded and have algorithms in place to deal with every conceivable variable. The question is whether we tune those algorithms to prioritise the safety of the passengers over other people.
Yes, either way it will be much safer, but there are still questions to be answered. You're saying something like "it doesn't matter whether candy or vegetables are more healthy because at least I'm not starving."
It blows my mind that people would rather drive a car themselves than be driven by one that could sacrifice them when they know the latter is categorically VASTLY safer
Personally, I think that the human 'driver' should be given the option to program what the car should do in this scenario. That way, the driver is still given the choice, as it would be the case today, and the driver can face the consequence he/she has made
Keep Summer safe.
none, an autonomous car should never find itself in a situation where there is a chance for an accident. If you wider how, go and study the entire road law and you'll notice that they have thought of everything.
The driver should be the one sacrificed, imo. The driver has the best chance of survival: the seatbelt and airbags and the driver holds the most blame, such as failing to upkeep maintenance on the car.
Self driving cars should always be over-rideable. And you should need someone with a license in it always, if there is a malfunction with anything be it the brakes/steering/ or even the software which makes these decisions, the car should hand over control to a driver
This is very interesting, it is very important to determine what moral decisions should be made. I'm all for driverless cars and am eager to find out the result. What do you guys think about it?
Nooter I think the buyer should make the decision of weather they would want the car to save them in the case that another would be hurt or have them possibly get severely hurt in exchange for another
The car should probably make no decision. If it has to deal with more factor during traffic it only increases the chance of an accident occurring. Tell it to avoid obstacles (by the lack of detail you know how good my programming skills are) and other very basic commands. That should minimize casualties. Who is to blame if an accident occurs? That's more difficult to say. I think that depends on how the accident happened. If the car was hacked, the people who dealt with security. Faulty car, the manufacturers. If the driver was the cause, the driver.
1. Decide by law that a group of people will be a less likely target.
2. Lone-standing people on the sides would know that and should have the habit to group together/run to the next group.
3. Cars which lost control will need a redundant system which will light up their whole front and show the decisions of the AI (aka a big red arrow to where to move).
4. Use external inflatable airbags before hitting a passenger (frontal hits are rather rare - most hits can be less dangerous if a person is pushed aside).
5. Let the software learn - each car is contributing to the system. If one unit learns how to handle a situation, ALL cars in the world will be able to do the same with the next update. This is a very huge thing because most driving errors are very primitive and repetitive.
The car should make the decision that will cost the fewer casualties
For sure, this is the better decision for the car and us too. No casualties are better!
If only all the TED presenters had this guy's skill... all the more impressive because English is his second language. Rock on, Iyad.
If you're in a Tesla, hitting the wall will likely not cause any deaths because their safety rating is so good. Seriously. People driving off of cliffs, getting ran over by semi's, and the driver of the Tesla walks out of the car unscathed.
THE COMPUTER NEEDS TO TAKE FAULT INTO ACCOUNT. If 2 pedestrians are jaywalking, a car shouldn't sacrifice the driver for a mistake the pedestrians made. Likewise, the pedestrians in the example all step onto the road right as a car is coming through. If the brakes really stop working, that is the driver's fault for not maintaining the vehicle, and therefore the driver should be sacrificed.
If there is a lack of maintenance, then that is the cars fault. It shouldn't drive if it is past its designated maintenance check, if the failure is a random catastrophic failure, then the part manufacturer is at fault and therefore the owner can never be to blame for mechanical failures.
The major problem here is not the actual discussion. It is how hack safe it would be
In case of any failure, alarms are sounded and if pedestrians move and clear out path then good, otherwise inform passengers to go into brace position and open airbags and then crash the car at an angle leading to minimum damage. If the safety systems are developed well in future, then crash might not turn out to be fatal.
if you eliminate human error i would argue you should also eliminate the drivers responsibility , if the software decides where the car goes then it should be also held accountable !
That's really interesting topic.
Right now most accidents are caused either by 1) driver or 2) car, while 2) might be prevented thanks to driver's reaction (he can stop driving if something isn't working)
with driveless cars the accidents could be caused mostly only by something not working in car, which would eliminate a great deal of accidents.
Pretty sure that situation @ 1:19 should never happen with driveless car, as it should be programmed to know where people can pass roads & slow before those places accordingly.
What we need then, is to add law that only allows to pass roads on marked places & increase number of those places.
If people would follow the law, there would be no accidents (unless of course something in car isn't working)
If a car has time to make a decision, it has had the time to have calculated all possible scenarios and modified its velocity so as to have avoided all dangerous possibilities. Accidents are the result of inattention, slow reactions and drivers not looking for possible dangers, the whole point of driverless cars is to calculate all possibilities all the time. In the rare case of an accident then the driverless car will not have had the possible danger calculated, if so it doesn't have the information to make a "moral" decision anyway.
If its a mechanical failure, then the car won't have the ability to effect any "moral" decisions it makes.
Different value for people for different people.
I think it should be passengers in the cars safety above all.
yeah ok, you can get in the suicide car that you dont even control then.
Hirvieläin What kind of argument is that? Who are you to decide what someone else think?
Let's get real, the auto companies are going to put their costumers first, if the driver is killed, it would be bad for business.
Then maybe we shouldn't let auto company make that decision.
Erikulum Thankfully we live in a democracy and not a socialist state. I doubt people would vote to pass a law that forces people to get in cars that may kill them in any given moment. By design or computer fault
But who will be held responsible if a driverless car is involved in an accident with or without human victims?
A driverless car is not the only and by far not the easiest way of reducing mortality on the roads. The number of mortal accidents per 100K cars differs dramatically between countries. From 6,8 in Germany to 104 in China or 130 in India (2013 data from Wikipedia). By reducing this rate in 5 countries only (China, India, Brazil, USA and Russia) to German level, we might reduce the global number of fatalities by 36% or save 450K lives per year. For that one needs to fight corruption, introduce more profound education for getting driver license, improve roads infrastructure and increase fines. If one has ever driven in Germany and in Russia (sorry, I've never been to China or India) one easily understands that before fixing ethical issues for AI there is looooots work to be done with human ethics in many countries.
These situations are a bit simplistic. For example, a car has a horn. People crossing the street have personal awareness and a desire to survive. If the car is making the moral decision then it knows there is a problem. It can audibly warn of the danger and split the liability between it's ability to maneuver and the pedestrian ability to move out of the way.
Just look at old footage of rally racing through crowds of people.
For a driverless car, control in an emergency should be given back to the user so that they can make the choice but more similar to the new 'fly by wire' where what the driver needs the on board computer makes possible by what ever means. You can't exactly put a car on trial for vehicular manslaughter...
There are common traffic rules. For example, if there is a sudden obstacle ahead, the driver should brake. So, why not simply obey such rules. Emergency braking reduce kinetic energy of car and prevent severe damage. Consequences of this action are quite well predictable. There are also people trying to commit suicide on dark street and there is a passenger in the car, which want a good life. If emergency braking don't help, bad luck.
When thinking about AI driven cars, what I call Robot Cars, I've always wondered about something similar. Let's say I'm being driven by robot car down the interstate at 75 mph and a deer runs in front of car from an undetectable position. Now if the car does nothing the car probably will get totaled when it hits the deer and it could seriously hurt or kill the driver and passengers. But if it swerves sufficiently to avoid the deer at such a high speed the risk of rolling over the car or otherwise wrecking is much higher. What would robot car do?
Or what if if it must choose between doing nothing and hitting the deer or swerving and likely hitting another vehicle in the other lane.
I assume it would slam on breaks as fast and hard as it could to minimize the impact...
Anthony Goodley practical cars aren't likely to roll over..... Just your statusmobile....
Anthony Goodley the computer and sensors of the car are going to be infinitely faster than any human could ever react. Deer on the road will be the type of thing that driverless cars will essentially eliminate ALL worry over.
You're not supposed to swerve to avoid deer, braking is the safest for everybody, even if there is some impact at front or rear. Swerving opens you up to flipping and multiple car collisions from cars going the opposite direction (and so have increased force on impact).
Ashley Feawen Our highways are divided here so little chance hitting an oncoming car. I've swerved many times and avoided deer successfully. Only hit a deer once. At 75 or 80 mph even a small deer will total your vehicle I learned. Insurance doesn't cover it here unless you have the expensive Comprehensive add on that isn't worth it.
Those results looked a lot like my results on that website.
Personally I think the risk from a driverless car making ethical decisions is too high. Not because its decision could be wrong or questioned but because people look at the driver to work out what a car is going to do. This is the same as walking through a crowd and no one crashing into each other. There is no driver to look to and so people must guess as to what the car is going to do. Someone could dive in front of the car because they didn't anticipate that it would swerve that way. If the car stayed in it's own lane then it is predictable. This gives people more chance to get out of the way as they know what the car will do.
Another issue is that when a person steps into the road they are taking a risk and should look, listen etc before crossing. A person on the pavement has not volunteered for this risk and should not be put at risk just because they happen to be in way of a car that wants to drive on the pavement.
The car should do what a human would do. Instinctively we would try not to hit what crosses our path. Meanwhile, in the back of our minds, we have our own survival instinct. We would quickly look all around and decide what to do. We would most likely slam on the brakes and pray that the person behind us is aware of the situation. But we all know that communication between vehicles is the first priority and will come first. So the car would tell the other cars that it is going to come to a stop and all other vehicles will act accordingly. Ultimately the passenger has the highest probability to survive the crash compared to the pedestrian.
It'll be a money thing. Why don't all cars come standard with ABS? It costs little enough nowadays it should be law like seatbelts.
So you think the only way to avoid accidents is swerving wildly or hitting the brakes?
You must be a wonderful driver.
We should make a game based on our traffic system so that we can have supercomputers figure out more advanced ways to deal with this moral issues by having it play it self over and over until mastering these issues.
I've never heard someone call it "department of transport"
I feel so unconfidable every time I hear ho people vote on ethical questions. So many egoistic and irational people.
Don't think people will need to buy the cars. Uber, for example, seems poised for self driving cars. They will most likely use the utility principle due to public pressure and passengers will just go along with it for the convenience. Prioritising immediate convenience over potential future negatives is something modern society specialises in. E.g. big data. The interesting point will come when such an accident occurs and the media covers it properly. I predict a huge outcry followed by no change when everyone loses interest due to the lack of easy solutions.
GREAT JOB MATE, GREAT ONE!!! AND GOOOD THOUGHTS ON THE ROBO LOWS
The car should take the action that the driver prefers and is within given legal limits. If some people want to give additional weight to saving others, saving children or saving themselves, they should be able to. In effect, this gives them the same choice they do now, only now its an explicit choice instead of a split-second and mostly instinctive decision.
With regards to PD type decisions, one could simply not make too sub-optimal decisions available. Additionally, the cars may take other drivers/people's decisions into account when deciding to solve an accident problem. They may give priority to people who have given actions that are less sefless when victims are going to be made, motivating people to take collective positive action.
Well just don't let it have a malfunction...
If the driverless cars are so safe in my opinion, it should be programed to save the most lives as possible. I for one do not want to have the job that decides who would live and die. But to answer the question I would have to say I would sacrifice myself instead of taking anyone else's life.
couldn't the car instead use the transmission to slow down and stall out?
Just 1 second more...
Is it 3:36 yet?
Show me a driving test that asks "Who would you hit," and I'll stop laughing at this argument about self-driving cars.
Why are we saying driverless cars can't have manual steering incorporated?
Because soon driver less cars will be a direct improvement on a human driver. The only reason these ethical debates are possible is that a computer has time to make a decision while a human will just act on instinct.
Since the owners of the driverless car are ultimately responsible for the cars actions, then the morality should be customizable. There, that was easy.
Self driving death machine!!! Maybe there should be some control over the options at least.
If you choose to make a self driving death machine, then you are going to go to jail or get the death penalty in the exact same way if you were the driver, so there is no point in proactively regulating the industry. We already have a justice system in place, let's use it before we resort to authoritarian regulations.
yeah that's what I was thinking . If it's your car it should be upon you to decide that what your car should do .
Wait, woudn´t willfully deciding to endanger other lives already be against the law? (might differ from country to country tough)
This is years and years and years from now, when it comes right down to it, self driving will be a feature just like cruise control, used only in specific driving situations, all cars running on their own all the time on all the roads - AIN'T GOING TO HAPPEN ANY TIME SOON IF EVER.
hello from 2024 with driverless cars
The cars will operate within the speed limits and if it crashed in a evasive situation, the air bags, seat belts and the safety cell will save the occupants,.so why the discussion.
Could someone provide the link to that website?
Is this a repost?
but realistically, passengers r more protected than the pedestrians.
The regulations are gonna have to be broader than just what the cars do or don't do. Example: Where I live crosswalks are often less than 8 seconds, enough for an able bodied adult to sprint across, but impossible to make if you're using a walker, wheelchair or heck just walking at a comfortable pace. If illegal jay walking suddenly makes you acceptable car fodder...
The car should prioritize the passenger's life first and foremost.
Why would you ever get into a driver-less vehicle that didn't have that in place?
If the driverless car are gonna save 90% of the people in the graph, it doesn't matter if is gonna decide to kill the passengers or the pedestrians. This decision doesn't impact most of the car crashes. Reckless drunk drives do. Build the one that people will love to buy and just then discuss the 10%.
Iyad Rahwan is not clearing thinking through to have concluded that engineers and programmers would had safety measures of what 'can' go wrong if any mechanical failure there would be a secondary measure to bring the vehicle to a stop and this would include smart systems but also there would be in place secondary backup systems to be automatically initiated.
I do have an example of my personal vehicle being a 2001 Saab Aero 2.3T sports sedan having a smart system and engine diagnosis computer w/ many sensors - should there be a serious or moderate engine problem detected as is still able to drive it but will be under powered that is indicated by dash lighted called 'Limp Home' mode to either travel back home or to the nearest automotive service facility or my trusted car service garage I go to for repairs.
If there is a more serious problem an cannot be advisable for 'Limp Home' mode as will indicate that on the dash or by your own judgement then you would shut off the engine and have parked on the safe side of the road or parking area and call for a tow truck and take it to your choice of servicing station.
So whether it is a smart safety system or being driver controlled that would have a direct manual or voice command to tell the vehicle to stop should any emergency arises or due to failures.
Just make one ways. ALOT of one ways. Or cars that can hop over accidents and glide to safety
The thing with the first issue he proposes is that a driverless car programmed in advance will be able to make a far better informed decision than a driver in a split second
It doesn't really matter what moral decision a self driving car should make, even remotely. It's a consumer product. Why would any consumer purchase a car that is going to save other people's lives over that of their own? That car is not going to be a remotely easy sell. Most people value their own life over that of any number of strangers.
You know, these scenarios are very unlikely. in the worst case scenario the engine can shut down and the care can put its brakes on.
that pretty much ignores the point of - why would a person buy a car that would even (in a VERY small % of a chance) choose the life of a stranger over their own life.. they paid for the car, the car should protect them over anyone else. Logically.
if you get caught in an accident you may die anyways. Regardless of whether a manually operated car gets in an accident or a driverless car gets in an accident, the passengers of either car may die. When you get in an accident, you have to rely on safety features from the car, that is what is supposed to protect you from death. The point of a driverless car is to minimize harm to people in the unfortunate scenario of getting in an accident.
They will certainly value the lives of their family in the car over strangers.
A car that can react instantly to a dangerous scenario can make decisions that minimize the scale and tragedy of the accident thus protecting your family. Much better than getting in a head on collision on a swervy mountainous road. I live in an area where this happens all the time.
Nothing exciting about not being able to drive your own vehicle. Last I remember people want to be in the drivers seat not a passenger in their own vehicle.
There will always be people do want to drive their own cars of course, but at the same time driving is just a method of transport and it's not much different to having a chauffeur drive you to wherever you need to be. Also driverless cars probably will be able to get you where you need to be faster, and I'm sure many people would love that...
Well, I see a need for it but particularly for the elderly, the disabled, the drunken, mentally challenged, or people who are having a bad day/going ballistic, and people with bad driving records. However, the driving experience should not be taken away, it would be a tragedy.
The manual driving feature should be available and accessible to those who are responsible competent drivers.
You are right though, it could be faster to reach to destinations because the decision making process for acceleration and deceleration would be accurately executed.
I guess if you lived through a the accident insurance company wouldn't pay out,because you had chosen the settings on your driverless car
wind directing vs power lines so an ejection seat ... again no AI CAN PREDICT FUTURE!
Inside the car you have an airbag outside you just have a bumper going to have to factor that in alsoI personally would swerve away from the people 100% of the time
This is a relatively small moral dilemma. Design decisions made in the production of that suit he is wearing may have had more significant life or death consequences than this programming moral dilemma, with less thought put into them. The road that the car is driving on? Almost certainly.
Is this a re-upload I feel like I've seen this before...
no, you must be from the future
Im confused imannuel
Kant wouldnt kill anyone so what would be the answer?
A solution would be to educate people to own their opinion and face the consequences of it. What kind of amoral child-mind chooses moral principles and want them applied only to others? Preschool children do that. Let's educate people so that they get out of preschool in that department and we'll have the over-all safest cars.
I *would* buy a car that could potentially sacrifice me in order not to harm others. Why wouldn't you if you're utilitarian? Of course we care about our life, but that's irrelevant because everyone cares about his/her life.
And if you don't want a machine to take moral decisions for you, then don't support driver-less cars at all.
I think most people would buy the coded to minimise damage car given the law that all driverless cars have to be coded that way. Because the choice would then be between driverless and driven, and driverless beats driven in safety everytime. It's a simple matter of implementing a law, like every other law. Like the law against prevention of free speech. Most people are annoyed with what the opposing political party says, but there is a law for freedom of speech we all generally respect and are punished for when we don't. There's no unsolved riddle here.
If the crash can't be avoided by perfect driving and instant braking by the driverless car then its the pedestrian's fault and the car should not swerve at all.
*(Inspector Gadget)* : Go go gadget car-helicopter!
I think that ultimately this becomes a political problem because the manufacturers will need to be told how to handle things. It might be by regulations, but I expect legislation. Either way, this brings the car companies and profit into the equation. Given that the management is duty bound to maximize profit for the shareholders, you have them making a utilitarian calculation of risk/benefit on their part. One thing we do know is plenty of politician pockets will be stuff with some serious $$$.
I would be most interested to hear how Elon Musk would answer these questions.
Sigh. Yes, people are too expensive. Long term robots are cheaper.
This talk misses the point of regulation. It's the roads, the vehicles and the drivers that are subject to regulation.
So for example shifting the option towards automation by scaling insurance fees for manual vehicles and changing regulation of existing roads to allow only driverless vehicles will go a long way towards public safety.
Also connected vehicles will have the ability to act collectively and intentionally to modify interactions with pedestrians or other situations like an accident.
The increased risk is very small no? Perhaps presenting the increased risk of what .005% chance u will die in 5 years of driving the car as that number will ease people's worries?
An evil friend of mine says that cars should be linked to the personal data of passers-by in the cloud and weighted, for example, between a 10-year-old and a 50-year-old to save a 10-year-old, between a dropout and a college student to save a college student, and between a criminal and a good citizen to save a good citizen. What do you think?
Well, a few years ago I would have been nice to say to my parents car "Alie, please drive me home. I am so wasted I can't even walk a straight line anymore."
I don't know if I'm thinking it in a too simple and naive way, but the car should crash into a wall.
If cars are designed to crash into a wall when they fail, the engineers could think a way of deploying an efficient, reliable and indipendent device that saves your life. On the contrary, if cars will kill pedestrians instead, engineers are trapped, they couldn't know what's the response of the pedestrian and the conception of a life-saving device for them is a lot more difficult.
In any case, the car should kill the driver anyway because when you decide to take the car to go wherever you want you inherently accept the fact that you might die in a lot more ways than a pedestrian. Also, you are "driving" a car, a massive thing that weights two tons and goes at 100 mph: you're the risk, not the pedestrians, and it's not fair that they have to suffer from your decisions
Alessandro Murtas what if there isn’t a wall. What if there is a line of school kids on one pavement and a line of church goers in the other.
The answer (which keeps being repeated) is not to somehow magically solve the given scenario but to place the car into a hypothetical scenario where injury or death is pretty much unavaoidable. Only at that point do you force the car to make a decision.
Oliver Foggin Well, frankly I did not think it that way. Thank you
I feel bad for Asimov. The man was a passionate genius and polymath who wrote or edited over 500 books but what everyone remembers him for is his hypothetical three laws of robotics.
شكرا على الترجمه
Deserves a standing ovation
I must be able to buy this type of car so I can make money....
we can not let one company monopolize the sistem.
"These cars will reduce deaths by cars by millions" "we should worry about the ethical ramifications tho..."
yeah this is a fun "exercise" but if automated cars save millions of lives from human driving error, this exercise is pointless. We shouldn't be worrying about the ethics of any single scenario. We should be worried about how many lives will be logically saved by less human driving mistakes.
I'm fairly sure that he briefly addressed this issue. I agree with you to some extent.
Just trust in people. People are lazy. If they can watch TV on the way to work, these cars are sold already.
I love driving. Why in the world would I want to be chauffeured around by a toaster?
Good, but most people would enjoy the extra time to sleep before work.
Or the car could just stop....
Let's get real, the auto companies are going to put their costumers first, if the driver is killed, it would be bad for business.
What about hackers? One smart-badass teenager with internet connection can play GTA with mi car?
They should each obey the rule of the selfish gene and try to protect themselves and their own occupants.
We cant have flying cars like we were condition to believe in the past would happen but we can have DRIVERLESS cars.........
ok whatever